Originally Posted by Durquavian
If anyone would pay attention to my posts.... Here is a point I have tried to make again and again. For a while I was getting annoyed at the "Intel is better/best/god" argument. But then it dawned on me. Most, and most probably means 95% or greater, of these programs were engineered and tested with Intel CPUs. So these benches test Intels capabilities far better than AMD. And with AMD being a far smaller share of the market, why would they? So what would happen if they setup a test that would be fairer, say test both to see how they actually worked against equally. HWBOT Prime is a new benchmark that does just that. And your claim that Intel is so much better than AMD or as you put it "a 5ghz 8350 is still slower than a stock 3930k", well AMD FX8350 is stomping all Intels, with the exception of a 32core Xeon (spelling there?). I already know why, but I'd rather get the Intel fanboys fuming first. And there is a reason why that will really make a lot of you AMD haters eat crow.
The most useful benchmarks, from anything resembling a practical standpoint, are those derived from, or representative of, real world applications.
Yes, some real world applications are written and compiled in a manner that favors Intel CPUs, and yes, sometimes AMD CPUs are not optimized for as well as they could be. However, this does not change the fact that these programs do actually run faster on Intel, and unless you have the ability to recompile or patch them yourself (which can be difficult for many proprietary apps) you get what you get.
The reference to HWBot Prime is absurd. It's as synthetic as it gets, and is an extremely new benchmark with a total of 117 submissions. Also, while open source, only the default build can submit scores, so it's currently stuck with whatever optimizations the creator used when compiling it, and thus susceptible to the same biases you are complaining about.